Industrial equipment anomaly detection optimization method and device based on multi-source data driving, electronic device and readable storage medium

Abstract

The application provides an industrial equipment anomaly detection optimization method and device based on multi-source data driving, electronic equipment and readable storage medium, device detection data of an industrial equipment and multi-source monitoring system metadata corresponding to the device detection data are acquired, the monitoring system is self-diagnosed based on the multi-source monitoring system metadata, and the health degree of the monitoring system is obtained. According to the health degree of the monitoring system, the device detection data is assigned a credibility weight. Combined with the credibility weight and the device detection data, the anomaly analysis result of the industrial equipment is obtained. In the scheme, the health state of the monitoring system is quantified and introduced into the device detection data analysis process, the misjudgment caused by the anomaly of the monitoring system is avoided, and the overall reliability and intelligent level of the industrial detection system are improved.

Description

Technical Field

[0001] This invention relates to the field of industrial inspection technology, and more specifically, to an optimized method, apparatus, electronic device, and readable storage medium for industrial equipment anomaly detection based on multi-source data-driven approaches. Background Technology

[0002] Current industrial monitoring systems generally adopt a "device-centric" monitoring model. Their technical architecture focuses on equipment status perception and mainly includes a data acquisition and transmission layer, a data analysis layer, and an application layer. The data acquisition and transmission layer collects physical signals from the equipment (such as vibration spectra and temperature curves) through vibration, temperature, and rotation speed sensors. The acquisition station uploads data via 4G / 5G / Wi-Fi links. The data analysis layer stores and processes the received equipment status data. Analysis methods are mostly based on threshold-based alarms (such as vibration amplitude exceeding limits) or simple machine learning models (such as regression and classification models) to predict trends or identify faults for single indicators. The application layer provides functions such as alarms for abnormal indicators, maintenance work orders, and data visualization of stored data.

[0003] The core of existing industrial monitoring systems is to monitor the health status of industrial equipment itself. However, these systems treat the monitoring system (sensors, data acquisition stations, and networks) as a completely reliable black box, lacking the ability to perceive and manage its own health status. Because monitoring systems may malfunction, this can lead to problems such as false alarms, missed alarms, and abnormal visualization. Summary of the Invention

[0004] The purpose of this invention is to provide an optimized method, apparatus, electronic device, and readable storage medium for industrial equipment anomaly detection based on multi-source data, so as to improve the overall reliability and intelligence level of industrial detection systems.

[0005] In a first aspect, the present invention provides an optimization method for industrial equipment anomaly detection based on multi-source data, the method comprising: Acquire equipment testing data of industrial equipment and metadata of the multi-source monitoring system corresponding to the equipment testing data; The monitoring system performs self-diagnosis based on the metadata of the multi-source monitoring system to obtain the health status of the monitoring system. Assign credibility weights to the device detection data based on the health status of the monitoring system; By combining the credibility weight and the equipment detection data, the anomaly analysis results of the industrial equipment are obtained.

[0006] In an optional implementation, the monitoring system includes a data acquisition station, a platform, and multiple sensors; The metadata of the multi-source monitoring system includes sensing layer metadata, transmission link metadata, and platform resource metadata.

[0007] In an optional implementation, the sensing layer metadata includes the battery voltage, current, internal resistance, power consumption, signal strength, number of communication requests, acquisition time, and calculation time of each sensor. The transmission link metadata includes data transmission delay, delay jitter, retransmission count, and signal quality parameters from each sensor to the acquisition station and from the acquisition station to the platform; The platform resource metadata includes CPU utilization, memory utilization, remaining disk space, message queue backlog, and number of services alive on the platform.

[0008] In an optional implementation, the step of performing self-diagnosis of the monitoring system based on the metadata of the multi-source monitoring system to obtain the health status of the monitoring system includes: The health of the perception layer is obtained based on the perception layer metadata, the health of the transmission link is obtained based on the transmission link metadata, and the health of the platform layer is obtained based on the platform resource metadata. Weighting coefficients are assigned to the health of the perception layer, the health of the transmission link, and the health of the platform layer, respectively. The health scores of the perception layer, transmission link, and platform layer are weighted and fused according to their respective weight coefficients to obtain the health score of the monitoring system.

[0009] In an optional implementation, the step of obtaining the health status of the perception layer based on the perception layer metadata includes: Collect standard metadata of the perception layer, and delineate anomaly boundaries based on the standard metadata of the perception layer; The system detects whether the metadata of the perception layer exceeds the abnormal boundary. If it exceeds the abnormal boundary, the health of the perception layer is determined from a preset abnormal range. If it does not exceed the abnormal boundary, the health of the perception layer is determined from a preset normal range.

[0010] In an optional implementation, the step of assigning confidence weights to the device detection data based on the health status of the monitoring system includes: The health status of the monitoring system is compared with a first preset threshold and a second preset threshold, wherein the first preset threshold is greater than the second preset threshold. If the health status of the monitoring system is greater than or equal to the first preset threshold, the credibility weight of the device detection data is set to the preset maximum weight. When the health status of the monitoring system is between the second preset threshold and the first preset threshold, the corresponding confidence weight is obtained based on the health status mapping of the monitoring system. If the health status of the monitoring system is less than or equal to the second preset threshold, the reliability weight of the device detection data is set to the preset minimum weight.

[0011] In an optional implementation, the step of combining the confidence weight and the equipment detection data to obtain the anomaly analysis result of the industrial equipment includes: The anomaly detection conditions are modified based on the aforementioned credibility weights; Based on the revised anomaly detection conditions and the equipment detection data, a preliminary analysis result of the equipment status is obtained; The preliminary analysis results are corrected based on the aforementioned credibility weights to obtain the final anomaly analysis results.

[0012] Secondly, the present invention provides an industrial equipment anomaly detection optimization device based on multi-source data-driven methods, the device comprising: The acquisition module is used to acquire equipment testing data of industrial equipment and metadata of the multi-source monitoring system corresponding to the equipment testing data; The self-diagnosis module is used to perform self-diagnosis of the monitoring system based on the metadata of the multi-source monitoring system to obtain the health status of the monitoring system. The allocation module is used to assign credibility weights to the device detection data based on the health status of the monitoring system. The analysis module is used to combine the confidence weight and the equipment detection data to obtain the anomaly analysis results of the industrial equipment.

[0013] Thirdly, the present invention provides an electronic device including a processor and a memory, the memory storing computer-executable instructions executable by the processor, the processor executing the computer-executable instructions to implement the method described in any of the foregoing embodiments.

[0014] Fourthly, the present invention provides a computer-readable storage medium storing computer-executable instructions, which, when invoked and executed by a processor, cause the processor to implement the method described in any of the foregoing embodiments.

[0015] Compared to existing technologies, this invention provides an optimized method, apparatus, electronic device, and readable storage medium for industrial equipment anomaly detection based on multi-source data. It acquires equipment detection data from industrial equipment and corresponding multi-source monitoring system metadata. Based on this metadata, it performs self-diagnosis of the monitoring system to obtain its health status. Confidence weights are assigned to the equipment detection data according to the system's health status. Combining these weights with the equipment detection data yields the anomaly analysis results for the industrial equipment. This solution quantifies the health status of the monitoring system and incorporates it into the equipment detection data analysis process, avoiding misjudgments caused by system anomalies and improving the overall reliability and intelligence level of the industrial detection system. Attached Figure Description

[0016] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments of the present invention will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 A flowchart illustrating the optimization method for industrial equipment anomaly detection based on multi-source data provided in this embodiment of the invention; Figure 2 This is a schematic diagram of the architecture of an industrial testing system provided in an embodiment of the present invention; Figure 3 for Figure 1 A flowchart of the sub-steps included in S12; Figure 4 for Figure 1 A flowchart of the sub-steps included in S13; Figure 5 for Figure 1 A flowchart of the sub-steps included in S14; Figure 6 A functional block diagram of an industrial equipment anomaly detection optimization device based on multi-source data provided in an embodiment of the present invention; Figure 7 This is a structural block diagram of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0018] The technical solutions of the present invention will now be described with reference to the accompanying drawings in the embodiments of the present invention.

[0019] Existing industrial monitoring systems lack the ability to assess the health status of the monitoring system itself. They only focus on the operating status of the monitored equipment, without systematically diagnosing the operating status of sensors, power supplies, communication links, computing platforms, and other components of the monitoring system. This makes it impossible to determine whether the equipment's monitoring data is affected by system anomalies. Consequently, existing technologies cannot distinguish between equipment anomalies and system anomalies. When abnormal equipment monitoring data occurs, existing systems cannot differentiate whether the anomaly originates from a genuine equipment malfunction or from system problems such as sensor undervoltage, degraded communication quality, or abnormal platform resources, easily leading to false alarms or missed alarms.

[0020] Furthermore, existing technologies suffer from the problem of unquantifiable reliability of equipment monitoring data. Current systems directly use collected equipment data as analysis input, lacking a quantitative evaluation mechanism for data reliability and validity. This leads to analysis results being treated the same even in system anomaly scenarios. In existing methods, the operation and maintenance model is passive and costly. System anomalies are typically discovered reactively after equipment data is abnormal or lost, requiring maintenance personnel to manually investigate the entire monitoring chain, resulting in low location efficiency and long response times.

[0021] To address the shortcomings of the existing technologies, this invention provides an optimized scheme for industrial equipment anomaly detection based on multi-source data. This scheme enables the monitoring system to perform self-diagnosis and quantifies the reliability of equipment detection data based on the diagnostic results. This avoids system anomalies from misleading equipment status analysis, alarms, and statistical results, thereby improving the overall reliability and intelligence level of the industrial detection system.

[0022] Please see Figure 1 The flowchart below shows the optimization method for industrial equipment anomaly detection based on multi-source data provided in this embodiment of the invention. The detailed steps of the optimization method for industrial equipment anomaly detection are described below.

[0023] S11, acquire equipment testing data of industrial equipment and metadata of the multi-source monitoring system corresponding to the equipment testing data; S12, perform self-diagnosis of the monitoring system based on the metadata of the multi-source monitoring system to obtain the health status of the monitoring system; S13, Assign credibility weights to the device detection data based on the health status of the monitoring system; S14, Combining the credibility weight and the equipment detection data, the anomaly analysis results of the industrial equipment are obtained.

[0024] The optimized method for detecting abnormalities in industrial equipment provided in this embodiment can be applied to industrial detection systems, such as... Figure 2As shown, the industrial detection system includes industrial equipment, a monitoring system, and an application layer. In addition, a self-diagnostic system is added. This self-diagnostic system can collect monitoring system metadata, analyze and process the monitoring system metadata, quantify the health status of the monitoring system, and obtain the health degree of the monitoring system.

[0025] The monitoring system can collect data from industrial equipment and obtain equipment testing data.

[0026] The quantified health status of the monitoring system reflects its reliability, which can be represented by assigning corresponding reliability weights to the collected equipment monitoring data. Finally, by combining the reliability weights and the equipment monitoring data, and following the established anomaly analysis method, the anomaly analysis results for the industrial equipment are obtained.

[0027] In this solution, the health status of the monitoring system is quantified by performing self-diagnosis on multi-source metadata of the monitoring system, and a credibility weight is assigned to the equipment detection data accordingly. This avoids misleading the equipment status analysis, alarm and statistical results due to system anomalies, thereby improving the overall reliability and intelligence level of the industrial detection system.

[0028] The specific implementation methods of each of the above steps will be explained in detail below.

[0029] In this embodiment, the monitoring system includes a data acquisition station, a platform, and multiple sensors, including but not limited to temperature sensors, vibration sensors, and rotation speed sensors.

[0030] Multiple sensors can be used to collect various equipment detection data from industrial equipment, including but not limited to temperature data, vibration data, and rotational speed data.

[0031] Meanwhile, the self-diagnostic system can collect multi-dimensional data from the monitoring system to obtain multi-source monitoring system metadata. Multi-source monitoring system metadata includes perception layer metadata, transmission link metadata, and platform resource metadata.

[0032] The sensing layer metadata includes, but is not limited to, the battery voltage, current, internal resistance, power consumption, signal strength, number of communication requests, acquisition time, and computation time of each sensor. The sensing layer metadata is data used to characterize the operating status of the sensors and the acquisition terminal.

[0033] The battery voltage of each sensor can be obtained by direct measurement or by reading from an analog-to-digital converter (ADC). A multimeter can be used to directly measure the voltage across the battery terminals. In embedded systems, the battery voltage can be read using an ADC.

[0034] The current of each sensor can be obtained by connecting an ammeter / sampling resistor in series. The ammeter is connected in series in the sensor circuit for measurement, or a known small resistor (sampling resistor) is connected in series in the circuit to measure the voltage drop across its two ends, and the current is calculated according to Ohm's law.

[0035] The internal resistance of each sensor can be obtained using specialized instruments or the AC injection method. The battery's internal resistance is very small and needs to be measured using a milliohm meter or a battery internal resistance tester. The DC discharge method (instantaneous high-current discharge to measure voltage drop) and the AC injection method (injecting AC signal to measure impedance) can be used.

[0036] The power consumption of each sensor can be calculated using the power consumption calculation formula after measuring the voltage and current. For digital sensors, their power consumption can also be estimated from the "operating current" and "operating voltage" parameters in the datasheet.

[0037] In addition, the sensor's signal strength can be obtained using system APIs or AT commands. For WiFi sensors, the signal strength value can be obtained through system APIs (such as Android's RConnectionMonitor). For cellular modules (NB-IoT / 4G), the signal strength can be queried by sending AT commands (such as AT+CSQ).

[0038] The number of communication requests from the sensor can be obtained using software techniques or system logs. A counter can be set in the sensor firmware or gateway program to accumulate the count each time a network connection is initiated, data is sent, or data is received. Alternatively, the number of communications can be counted by analyzing system logs or network packet captures (such as using Wireshark).

[0039] The sensor's data acquisition time can be obtained using code timing. Record the start timestamp at the point where the sensor begins data acquisition and the end timestamp at the point where acquisition is complete; the difference between the two timestamps is the acquisition time. High-precision timing functions such as System.currentTimeMillis() or System.nanoTime() can be used.

[0040] The computation time of sensors can be obtained through code timing or performance analysis, similar to data acquisition time, and timed before and after the execution of the algorithm or computation task. In Java, the StopWatch utility class can be used to more easily manage the time of multiple tasks. For complex systems, performance analysis tools or logging can be used for tracking.

[0041] The transmission link metadata includes, but is not limited to, data transmission delay, jitter, retransmission count, and signal quality parameters from each sensor to the acquisition station and from the acquisition station to the platform. The transmission link metadata is used to characterize the stability during data transmission.

[0042] Taking the transmission link from the sensor to the data acquisition station as an example, the data transmission delay can be addressed using clock synchronization, protocol analysis, or queuing models. Specifically, in the clock synchronization method, high-precision clock sources (such as GPS receivers) can be deployed at both the sensor and the data acquisition station, utilizing their pulse-per-second (PPS) signals for synchronization. A local timestamp is recorded when sending data packets, and the receiving end records its own timestamp upon receipt; the difference between the two is the transmission delay.

[0043] In protocol analysis, the round-trip time (RTT) is calculated by analyzing the time difference between request and response packets in a known communication protocol (such as Modbus TCP), and then the one-way delay is estimated.

[0044] In the queuing model method, nodes in wireless sensor networks are modeled as queuing systems, and theoretical delays are calculated by analyzing the queuing time, service time, and retransmission probability of data packets in the nodes.

[0045] In addition, latency jitter refers to the degree of variation in the transmission latency of consecutive data packets, reflecting network stability. It can be obtained by performing multiple consecutive latency measurements to obtain a series of latency samples and calculating the absolute value of the difference between adjacent samples. Alternatively, an oscilloscope or network analyzer can be used to capture and decode the signal at the physical layer, and the jitter performance of the transmission system can be directly measured by analyzing the jitter at the arrival time of a specific synchronization character.

[0046] The number of retransmissions reflects the reliability of data transmission and the degree of network congestion, and can be obtained through methods such as protocol stack statistics, application layer log analysis, or network packet capture analysis.

[0047] Signal quality parameters are the fundamental physical factors that affect the aforementioned delay, jitter, retransmission count, etc., and mainly include signal strength and link quality.

[0048] Platform resource metadata includes, but is not limited to, CPU utilization, memory utilization, remaining disk space, message queue backlog, and number of services alive on the platform. Platform resource metadata is data used to characterize the running status of the backend system.

[0049] Among them, CPU utilization, memory utilization, and remaining disk space in the platform are the most basic monitoring items, which can be obtained directly from the operating system level.

[0050] The message queue backlog can be obtained from the message queue middleware, for example, through API calls, client tools, or named queries.

[0051] The number of service liveness instances typically refers to the number of healthy instances in a specific service cluster, which can be obtained primarily through service discovery and health monitoring mechanisms.

[0052] Based on this, the monitoring system performs self-diagnosis using metadata from the multi-source monitoring system to obtain the system's health status. For details, please refer to [link / reference needed]. Figure 3 This step can be achieved in the following way: S121, obtain the health status of the perception layer based on the perception layer metadata, obtain the health status of the transmission link based on the transmission link metadata, and obtain the health status of the platform layer based on the platform resource metadata; S122, assign weight coefficients to the health of the perception layer, the health of the transmission link, and the health of the platform layer, respectively; S123, the health of the perception layer, the health of the transmission link, and the health of the platform layer are weighted and fused according to their respective weight coefficients to obtain the health of the monitoring system.

[0053] In this embodiment, the method of obtaining the health status of the perception layer based on the perception layer metadata can be achieved by, for example, by pre-training an evaluation model for health assessment, using the evaluation model and obtaining the health status of the perception layer based on the perception layer metadata, or by setting quantization rules for the data of each dimension in the perception layer metadata, obtaining the corresponding quantized value based on the quantization rules and the data values ​​of each dimension, and then averaging or weighting the quantized values ​​to obtain the health status of the perception layer.

[0054] In addition, as another possible implementation, the step of obtaining the health status of the perception layer based on the perception layer metadata can also be achieved in the following ways: Collect standard metadata of the perception layer and delineate the abnormal boundary based on the standard metadata of the perception layer; detect whether the metadata of the perception layer exceeds the abnormal boundary. If it exceeds the abnormal boundary, determine the health of the perception layer from the preset abnormal range. If it does not exceed the abnormal boundary, determine the health of the perception layer from the preset normal range.

[0055] In this embodiment, standard metadata for the perception layer can be pre-collected. This standard metadata can include multiple data points indicating that the monitoring system is in a normal and healthy state. By analyzing these multiple standard metadata points, the data range within which the perception layer metadata of the monitoring system is located under normal and healthy conditions can be determined. Therefore, anomaly boundaries can be defined based on the boundaries of this data range.

[0056] In the actual detection process, the obtained perception layer metadata is compared with the abnormal boundary. If the perception layer metadata exceeds the abnormal boundary, it indicates that the perception layer metadata is in an abnormal state. If the perception layer metadata does not exceed the abnormal boundary, it indicates that the perception layer metadata is in a normal state.

[0057] In addition, in this embodiment, a preset abnormal range and a preset normal range can be preset. The preset abnormal range may include multiple abnormal health levels, and the preset normal range may include multiple normal health levels.

[0058] If it is determined that the metadata of the perception layer is in an abnormal state, the health of the perception layer can be determined from multiple abnormal health values ​​within a preset abnormal range. For example, it can be selected randomly, or the corresponding abnormal perception value can be determined from multiple abnormal perception values ​​based on the degree to which the metadata of the perception layer exceeds the abnormal boundary.

[0059] Furthermore, if it is determined that the metadata of the perception layer is in a normal state, the health layer of the perception layer can be determined from multiple normal health levels within a preset normal range, and the determination method is similar to that described above.

[0060] Furthermore, the implementation methods for obtaining transmission link health based on transmission link metadata and platform layer health based on platform resource metadata are similar to those for obtaining perception layer health based on perception layer metadata, and will not be elaborated here.

[0061] After obtaining the health status of the perception layer, transmission link, and platform layer, weight coefficients are assigned to each of them. For example, the weight coefficients can be assigned based on the degree of influence of the perception layer health status, transmission link health status, and platform layer health status on the health status of the monitoring system. For instance, the perception layer health status mainly reflects the device status of the sensors and directly affects the reliability of the collected device detection data. Therefore, the weight coefficient of the perception layer health status can be set to a larger value.

[0062] In addition, in order to unify the standards of health across various dimensions, the values ​​of the perception layer health, transmission link health, and platform layer health can be normalized to a set range, such as between 0 and 1.

[0063] Based on this, the health scores of the perception layer, transmission link, and platform layer are weighted and fused according to the assigned weight coefficients to obtain the health score of the monitoring system, which can be characterized as follows: Hs = α·H_sensor + β·H_link + γ·H_platform Where Hs represents the health of the monitoring system, H_sensor represents the health of the perception layer, H_link represents the health of the transmission link, and H_platform represents the health of the platform layer. α, β, and γ are weighting coefficients, and α+β+γ=1.

[0064] Once the health status of the monitoring system is obtained, a reliability weight is assigned to the equipment detection data based on the health status of the monitoring system. For details, please refer to [link to relevant documentation]. Figure 4 This step can be achieved in the following ways: S131, compare the health status of the monitoring system with a first preset threshold and a second preset threshold respectively, wherein the first preset threshold is greater than the second preset threshold; S132, if the health status of the monitoring system is greater than or equal to the first preset threshold, the credibility weight of the device detection data is set to the preset maximum weight. S133, when the health of the monitoring system is between the second preset threshold and the first preset threshold, obtain the corresponding confidence weight based on the health mapping of the monitoring system; S134, if the health status of the monitoring system is less than or equal to the second preset threshold, the credibility weight of the device detection data is set to the preset minimum weight.

[0065] In this embodiment, the higher the health value of the monitoring system, the better the health status of the monitoring system; conversely, the lower the health value of the monitoring system, the worse the health status of the monitoring system.

[0066] Based on this, in this embodiment, a first preset threshold and a second preset threshold are pre-set as evaluation standards. If the health status of the monitoring system is greater than or equal to the first preset threshold, it indicates that the health status of the monitoring system is good, further indicating that the reliability of the equipment detection data obtained by the monitoring system is high. Therefore, the reliability weight of the equipment detection data can be set to a preset maximum weight. The higher the reliability weight of the equipment detection data, the higher the reliability of the equipment detection data.

[0067] If the health status of the monitoring system is between the first preset threshold and the second preset threshold, the corresponding confidence weight can be obtained by mapping the health status of the monitoring system in a linear or non-linear manner.

[0068] Furthermore, if the health status of the monitoring system is less than or equal to the second preset threshold, it indicates that the health status of the monitoring system is poor and the reliability of the equipment detection data obtained from the monitoring system is low. Therefore, the reliability weight of the equipment detection data can be set to the preset minimum weight.

[0069] Based on this, and combining credibility weights and equipment testing data, the anomaly analysis results for industrial equipment are obtained. For details, please refer to [link / reference needed]. Figure 5 This step can be achieved in the following way: S141, The anomaly judgment conditions are modified according to the credibility weight; S142, Based on the revised anomaly judgment conditions and the equipment detection data, a preliminary analysis result of the equipment status is obtained; S143, The preliminary analysis results are corrected based on the credibility weight to obtain the final anomaly analysis results.

[0070] In this embodiment, anomaly detection conditions may include anomaly alarm triggering conditions and alarm levels. A high confidence weight indicates high reliability of the device detection data. In this case, the original device anomaly alarm triggering conditions and alarm levels can be maintained. A low confidence weight indicates low reliability of the device detection data. In this case, the anomaly alarm triggering conditions can be modified to be more stringent. For example, for device detection data where a high value indicates anomaly, the threshold value in the triggering condition can be modified to a lower threshold; for device detection data where a low value indicates anomaly, the threshold value in the triggering condition can be modified to a higher threshold. If the reliability of the device detection data is low, the alarm level can be modified to a higher level. This can prevent missed detections due to poor health status of the monitoring system.

[0071] Based on this, preliminary analysis results of the equipment status are obtained based on the revised anomaly judgment conditions and equipment detection data, including whether an anomaly alarm triggering conditions and alarm levels.

[0072] Finally, the preliminary analysis results can be revised based on the confidence weight to obtain the final anomaly analysis results. For example, if the preliminary analysis results indicate that the anomaly alarm triggering condition has been met, but the confidence weight is too low, such as below the preset minimum threshold, it indicates that the device detection data at this time is completely unreliable. In this case, the anomaly alarms in the preliminary analysis results can be suppressed or frozen to avoid false alarms.

[0073] The multi-source data-driven optimization method for industrial equipment anomaly detection provided in this embodiment is the first to quantify the health status of the monitoring system and introduce it into the equipment detection data analysis process, avoiding misjudgments caused by system anomalies. It effectively distinguishes between equipment anomalies and system anomalies, improving alarm accuracy. Through a credibility weighting mechanism, the reliability of equipment health assessment and statistical analysis results is improved. This solution promotes a shift in operation and maintenance mode from passive response to proactive prevention, reducing operation and maintenance costs.

[0074] Based on the same inventive concept, please refer to Figure 6This invention also provides a functional module diagram of an industrial equipment anomaly detection optimization device driven by multi-source data. This embodiment can divide the functional modules of the industrial equipment anomaly detection optimization device driven by multi-source data according to the above method embodiment. For example, each function can be divided into its own functional modules, or two or more functions can be integrated into one processing module. The integrated modules can be implemented in hardware or as software functional modules. It should be noted that the module division in this embodiment is illustrative and only represents one logical functional division; other division methods may be used in actual implementation.

[0075] For example, when dividing functional modules according to their respective functions, Figure 6 The illustrated industrial equipment anomaly detection and optimization device based on multi-source data is only a schematic diagram. This device may include an acquisition module, a self-diagnosis module, an allocation module, and an analysis module. The functions of each module are described in detail below.

[0076] The acquisition module is used to acquire equipment testing data of industrial equipment and metadata of the multi-source monitoring system corresponding to the equipment testing data; The self-diagnosis module is used to perform self-diagnosis of the monitoring system based on the metadata of the multi-source monitoring system to obtain the health status of the monitoring system. The allocation module is used to assign credibility weights to the device detection data based on the health status of the monitoring system. The analysis module is used to combine the confidence weight and the equipment detection data to obtain the anomaly analysis results of the industrial equipment.

[0077] The industrial equipment anomaly detection optimization device based on multi-source data driven provided in this embodiment can be used to execute the industrial equipment anomaly detection optimization method based on multi-source data driven in any of the above embodiments. For details not covered in this embodiment, please refer to the corresponding descriptions in the above embodiments. This embodiment will not elaborate further here.

[0078] Please see Figure 7 This is a structural block diagram of an electronic device provided in an embodiment of the present invention. The electronic device can be a computer device, a server, etc. The aforementioned self-diagnostic system can be carried on this electronic device, which performs its related functions. The electronic device also includes a memory, a processor, and a communication module. The memory, processor, and communication module are electrically connected directly or indirectly to each other to achieve data transmission or interaction. For example, these components can be electrically connected to each other through one or more communication buses or signal lines.

[0079] The memory is used to store computer programs or data. Memory can be, but is not limited to, Random Access Memory (RAM), Read Only Memory (ROM), Programmable Read-Only Memory (PROM), Erasable Programmable Read-Only Memory (EPROM), Electrically Erasable Programmable Read-Only Memory (EEPROM), etc.

[0080] The processor is used to read / write data or programs stored in the memory and execute the industrial equipment anomaly detection optimization method based on multi-source data driven provided in any embodiment of the present invention.

[0081] The communication module is used to establish communication connections between electronic devices and other communication terminals via a network, and to send and receive data via the network.

[0082] It should be understood that, Figure 7 The structure shown is only a schematic diagram of an electronic device; the electronic device may also include components that are larger than those shown. Figure 7 The more or fewer components shown, or having the same Figure 7 The different configurations shown.

[0083] Furthermore, embodiments of the present invention also provide a computer-readable storage medium storing machine-executable instructions, which, when executed, implement the industrial equipment anomaly detection optimization method based on multi-source data driven provided in the above embodiments.

[0084] Specifically, the computer-readable storage medium can be a general-purpose storage medium, such as a removable disk or hard disk. When the computer program on the computer-readable storage medium is run, it can execute the aforementioned optimization method for industrial equipment anomaly detection based on multi-source data. The processes involved in the execution of the executable instructions on the computer-readable storage medium can be referred to the relevant descriptions in the above method embodiments, and will not be detailed here.

[0085] In the embodiments provided by this invention, it should be understood that the disclosed apparatus and method can be implemented in other ways. The apparatus embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. Furthermore, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Additionally, the displayed or discussed mutual couplings, direct couplings, or communication connections may be through some communication interfaces; indirect couplings or communication connections between devices or units may be electrical, mechanical, or other forms.

[0086] Furthermore, the units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0087] Furthermore, the functional modules in the various embodiments of the present invention can be integrated together to form an independent part, or each module can exist independently, or two or more modules can be integrated to form an independent part.

[0088] It should be noted that if the functionality is implemented as a software module and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0089] In this document, relational terms such as first and second are used only to distinguish one entity or operation from another entity or operation, without necessarily requiring or implying any such actual relationship or order between these entities or operations.

[0090] The above description is merely an embodiment of the present invention and is not intended to limit the scope of protection of the present invention. For those skilled in the art, the present invention can have various modifications and variations. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. An optimized method for anomaly detection in industrial equipment based on multi-source data, characterized in that, The method includes: Acquire equipment testing data of industrial equipment and metadata of the multi-source monitoring system corresponding to the equipment testing data; The monitoring system performs self-diagnosis based on the metadata of the multi-source monitoring system to obtain the health status of the monitoring system. Assign credibility weights to the device detection data based on the health status of the monitoring system; By combining the credibility weight and the equipment detection data, the anomaly analysis results of the industrial equipment are obtained.

2. The optimization method for industrial equipment anomaly detection based on multi-source data as described in claim 1, characterized in that, The monitoring system includes a data acquisition station, a platform, and multiple sensors; The metadata of the multi-source monitoring system includes sensing layer metadata, transmission link metadata, and platform resource metadata.

3. The optimization method for industrial equipment anomaly detection based on multi-source data as described in claim 2, characterized in that, The metadata of the perception layer includes the battery voltage, current, internal resistance, power consumption, signal strength, number of communication requests, acquisition time, and calculation time of each sensor. The transmission link metadata includes data transmission delay, delay jitter, retransmission count, and signal quality parameters from each sensor to the acquisition station and from the acquisition station to the platform; The platform resource metadata includes CPU utilization, memory utilization, remaining disk space, message queue backlog, and number of services alive on the platform.

4. The optimization method for industrial equipment anomaly detection based on multi-source data as described in claim 2, characterized in that, The step of performing self-diagnosis of the monitoring system based on the metadata of the multi-source monitoring system to obtain the health status of the monitoring system includes: The health of the perception layer is obtained based on the perception layer metadata, the health of the transmission link is obtained based on the transmission link metadata, and the health of the platform layer is obtained based on the platform resource metadata. Weighting coefficients are assigned to the health of the perception layer, the health of the transmission link, and the health of the platform layer, respectively. The health scores of the perception layer, transmission link, and platform layer are weighted and fused according to their respective weight coefficients to obtain the health score of the monitoring system.

5. The optimization method for industrial equipment anomaly detection based on multi-source data as described in claim 4, characterized in that, The step of obtaining the health status of the perception layer based on the perception layer metadata includes: Collect standard metadata of the perception layer, and delineate anomaly boundaries based on the standard metadata of the perception layer; The system detects whether the metadata of the perception layer exceeds the abnormal boundary. If it exceeds the abnormal boundary, the health of the perception layer is determined from a preset abnormal range. If it does not exceed the abnormal boundary, the health of the perception layer is determined from a preset normal range.

6. The optimization method for industrial equipment anomaly detection based on multi-source data as described in claim 1, characterized in that, The step of assigning credibility weights to the device detection data based on the health status of the monitoring system includes: The health status of the monitoring system is compared with a first preset threshold and a second preset threshold, wherein the first preset threshold is greater than the second preset threshold. If the health status of the monitoring system is greater than or equal to the first preset threshold, the credibility weight of the device detection data is set to the preset maximum weight. When the health status of the monitoring system is between the second preset threshold and the first preset threshold, the corresponding confidence weight is obtained based on the health status mapping of the monitoring system. If the health status of the monitoring system is less than or equal to the second preset threshold, the reliability weight of the device detection data is set to the preset minimum weight.

7. The optimization method for industrial equipment anomaly detection based on multi-source data as described in claim 1, characterized in that, The step of combining the confidence weight and the equipment detection data to obtain the anomaly analysis result of the industrial equipment includes: The anomaly detection conditions are modified based on the aforementioned credibility weights; Based on the revised anomaly detection conditions and the equipment detection data, a preliminary analysis result of the equipment status is obtained; The preliminary analysis results are corrected based on the aforementioned credibility weights to obtain the final anomaly analysis results.

8. An industrial equipment anomaly detection and optimization device based on multi-source data-driven methods, characterized in that, The device includes: The acquisition module is used to acquire equipment testing data of industrial equipment and metadata of the multi-source monitoring system corresponding to the equipment testing data; The self-diagnosis module is used to perform self-diagnosis of the monitoring system based on the metadata of the multi-source monitoring system to obtain the health status of the monitoring system. The allocation module is used to assign credibility weights to the device detection data based on the health status of the monitoring system. The analysis module is used to combine the confidence weight and the equipment detection data to obtain the anomaly analysis results of the industrial equipment.

9. An electronic device, characterized in that, The method includes a processor and a memory, the memory storing computer-executable instructions executable by the processor, the processor executing the computer-executable instructions to implement the method of any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions that, when invoked and executed by a processor, cause the processor to perform the method according to any one of claims 1 to 7.

Images